fix: parse imported and embedded tags like CLI arguments so hyphenated tags survive

orvanitag/orvani.py

@@ -1,5 +1,4 @@
 import os
-import re
 import json
 import shutil
 import tomllib
@@ -42,7 +41,6 @@ ERRORS:
 class OrvaniTag:
 
     ENABLE_PLUGINS = True # Toggle use of plugins
-    REGEX = r'\s*([@+]\w+|#\w+)'
 
     def __init__(self, path: str=CONTEXT_FILE, cache: str=CACHE_FILE, contexts=CONTEXTS, plugins:list=DEFAULT_PLUGINS):
         """
@@ -158,12 +156,14 @@ class OrvaniTag:
 
             # Add each argument list element to context dict based on the
             # prefix symbols defined in the schema
-            if arg[0] in self.SCHEMA.values():
+            tag = arg[1:].replace(',', '')
+            if arg[0] in self.SCHEMA.values() and tag:
                 for context_key in self.SCHEMA.keys():
                     if self.SCHEMA[context_key] != None and arg.startswith(self.SCHEMA[context_key]):
-                        context[context_key].append(arg[1:].replace(',', ''))
+                        if tag not in context[context_key]:
+                            context[context_key].append(tag)
             else:
-                # No prefix, so it is part of the note
+                # No prefix (or a bare prefix), so it is part of the note
                 note.append(arg)
 
         if note:
@@ -325,14 +325,10 @@ class OrvaniTag:
 
     def __get_argv(self, context_string: str):
         """
-        Make a list of args from a context string to pass as an *argv parameter
+        Make a list of args from a context string to pass as an *argv parameter.
+        Words are classified by __context, the same way as command line arguments.
         """
-
-        tags = re.findall(self.REGEX, context_string)
-        note = re.sub(self.REGEX, '', context_string).lstrip()
-        argv = ' '.join(tags).split(' ') + [note]
-        argv = [x for x in dict.fromkeys(argv) if x != ''] # make unique and remove empty items
-        return argv
+        return context_string.split()
 
 
     def add(self, file: str, *argv: str):

tests/test_orvani.py

@@ -243,3 +243,14 @@ def test_import_skips_known_tags_and_empty_lines(tree):
     (tree / 'context.txt').write_text('./a.mp3 --- #lofi #snare\n./b.mp3 --- \n')
     assert OrvaniTag().load('context.txt') == [0, 'context.txt', 1]
     assert lines(tree) == ['a.mp3 --- #lofi #snare']
+
+
+def test_import_keeps_hyphenated_tags(tree):
+    (tree / 'context.txt').write_text('./a.mp3 --- @home-office #lo-fi rock+roll take-2\n')
+    assert OrvaniTag().load('context.txt')[0] == 0
+    assert lines(tree) == ['a.mp3 --- @home-office #lo-fi rock+roll take-2']
+
+
+def test_duplicate_and_bare_prefixes(tree):
+    OrvaniTag().add('a.mp3', '#lofi', '#lofi', 'C', '+', 'D')
+    assert lines(tree) == ['a.mp3 --- #lofi C + D']

tests/test_plugins.py

@@ -151,3 +151,12 @@ def test_write_ignores_configured_types_the_plugin_cannot_handle(cwd):
     OrvaniTag().add('doc.abc', '#lofi')
     assert OrvaniTag().write_metadata('doc.abc', 'pdf') == [4, ['.pdf'], 'pdf', True]
     assert (cwd / 'doc.abc').read_text() == 'not a pdf'
+
+
+def test_pdf_keeps_hyphenated_tags(cwd):
+    writer = PdfWriter()
+    writer.add_blank_page(72, 72)
+    writer.add_metadata({'/Keywords': '+side-project #lo-fi'})
+    writer.write(cwd / 'doc.pdf')
+    OrvaniTag().read_metadata('doc.pdf')
+    assert (cwd / orvani.CONTEXT_FILE).read_text() == 'doc.pdf --- +side-project #lo-fi\n'